Remote Identity Validation Technology Demonstration | Homeland Security
U.S. flag

An official website of the United States government

  1. Science and Technology
  2. Expertise
  3. Technology Centers
  4. Biometric and Identity Technology Center
  5. Remote Identity Validation Technology Demonstration

Remote Identity Validation Technology Demonstration

The U.S. Department of Homeland Security (DHS) Science and Technology Directorate (S&T), in partnership with the Transportation Security Administration (TSA), the National Institute of Standards and Technology (NIST),  Homeland Security Investigations (HSI) Forensic Laboratory, and other federal agencies, will host a series of test events to challenge industry to deliver secure, accurate, and easy-to-use remote identity validation technologies to combat identity fraud when users apply for government services, open bank accounts, or verify social media accounts.  

S&T and its partners will evaluate the ability of systems to authenticate identity documents, assess the "liveness" of selfie photos, and evaluate identity verification using images taken with smartphones and similar commodity devices.

2024. Remote Identity Validation Technology Demonstration.

About the Technology Demonstrations

Through S&T’s Biometric and Identity Technology Center, these test events will enable industry to:

  • develop more secure, accurate, and easy to use technologies;
  • objectively measure performance against realistic and sophisticated attacks;
  • answer questions about the overall performance and risks of these technologies for use in commercial or government applications; and
  • inform efforts to standardize and certify technologies that are effective against sophisticated and rapidly evolving attacks.  

These test events are held at the Maryland Test Facility (MdTF), located in Upper Marlboro, Maryland. For more information, visit the MdTF RIVTD website.

Infographic with three sections titled "Overview of the Remote Identity Validation Technology Demonstration". Prepared by the IDSL for the DHS Science and Technology Directorate. The first section is titled Track 1: ID Validation. Two female figures stand over tables, each using a smartphone to photograph ID documents (ID’s shown above them). The ID on the left is labeled "Accept ID", in green. Below the driver's license is text: "Dataset of over 1,000 REAL state ID card photos. The ID on the right is labeled "Reject ID", in red and the ID reads "Fake ID". Below the fake ID, there is text that reads "Dataset of over 1,000 FAKE state ID card photos". The second section is titled "Track 2: Match to Document". Two figures stand, taking selfies. Each label has a selfie image of the person's face below it, connected to the same ID document. The ID’s owner is labeled "Verify Identity", and the other figure is labeled "Fail to Match". Next to the selfie images is text: "Dataset of selfie photos and genuine documents from over 1,000 people”. Next to the ID, there is text on the left side: "Over 1,000 mated comparisons" and on the right side "Over 500,000 non-mated comparisons". The third section is titled "Track 3: Presentation Attack Detection". Two figures stand taking selfies, and one of them is wearing a solid-colored mask. The selfies are shown, and the person without mask’s selfie is labeled “Accept Selfie". The person with a mask’s selfie is labeled “Detect Attack”. Text below the selfies reads "Tested with over 600 diverse bona fide users" for the left image and "Tested with over 1,200 presentation attacks" for the right image.

Technology Demonstration

S&T invites technology developers with demonstrated expertise in remote identity validation capabilities to participate in the Remote Identity Validation Technology Demonstration. RIVTD is conducted in three tracks and each track is demonstrated separately. 

  • Track 1: Assess the validity of an identity document (U.S. state-issued driver’s license or identification card)
  • Track 2: Match a “selfie” photo to the photo on the identity document
  • Track 3: Assess the “liveness” of the “selfie” photograph

All Tracks are currently closed for applications. Organizations interested in participating should email peoplescreening@hq.dhs.gov to be notified of future events.

Results

Infographic with three sections, titled "Remote Identity Validation Technology Demonstration - Document Validation". The first subsection is titled "Testing Document Validation Systems". It has four graphics. The first one is titled "Genuine and Fake IDs". Two people stand over two tables using smartphones to photograph ID documents. The ID documents are shown, and one is labeled "Fake ID". The genuine ID and photographer are labeled: "Document False Reject Rate (DFRR) > 1,000 REAL state ID card photos". The fake ID and photographer are labeled "Document False Accept Rate (DFAR) > 1,000 FAKE state ID card photos". The second graphic is titled "Multiple Smartphones" and shows three different smartphone models (A, B and C). The third graphic is titled "ID Documents Sampled Across Multiple States" and shows genuine and fake IDs connected to a box containing a map of the United States. The fourth graphic is titled "12 Document Validation Systems" and shows 12 computer servers. The second subsection is titled "Insights from Testing". An archery target has a label "Recommend selecting systems with error rates below 10%". The target is connected to a graphic of a paper document labeled "Adopted by NIST as the Benchmark for Digital Identity Systems, NIST 800-63-4". The second section is titled "What Impacts Document Validation System Performance?" and has two pairs of graphics (each with two graphics). Both pairs show a person photographing an ID document with a smartphone, one of which authenticates and one of which fails to authenticate. The first pair is titled "Validation Errors Depended on Document State of Issuance" and labeled "State A license validated" and "State B license error". The second pair of graphics is titled "Validation Errors Depended on the Smart Phone Used to Collect the Image" and labeled “Phone A scan validated” and “Phone B scan error”. The final graphic in the section has a warning symbol and is labeled "System deployers should ensure compatibility with their users' documents and devices." Prepared by the IDSL for the DHS Science and Technology Directorate.

 

Infographic with three sections, titled "Remote Identity Validation Technology Demonstration - Selfie Match to Document". The first section is titled "56% of Systems Matched Documents to Selfies Better than 99% of the Time". A selfie photograph of a person is compared to their ID, and the comparison succeeds. Nine of sixteen servers are green, and the remainder are gray. A green servers are labeled "The best system matched 100% of documents to selfies". A gray servers are labeled "The worst system matched 0% of documents to selfies". The second section is titled "61% of Systems Met or Exceeded the Target Security Level, Rejecting 99.99% of Random Imposters". A selfie photograph of a person is compared to a different person’s ID, and the comparison fails. Ten of sixteen servers are green, and the remainder six are gray. Another graphic is titled "Imposter Rejection Rate: Min: 90%, Median: 99.99%, Max: 99.999%". The graphic shows two selfies of different people being compared and is labeled "Demographically Matched Imposters are 10x More Successful". The third section is titled "What Causes Errors?" A person is photographing an ID document on a table with a smartphone. This graphic is labeled "Extracting Document Templates". This graphic connects to another set of graphics showing two different ID documents in multiple orientations. Both ID documents, a horizontal and vertical ID card, belong to the photographer. This graphic is titled "Matching Documents to Selfies". This graphic is connected to a stacked bar chart with one bar and "Errors" on the x axis. The bar chart is labeled "2% of errors due to selfie extraction, 43% of errors due to selfies not matching, 55% of error due to document extraction". Next to the bar chart is a graphic titled "5 Systems Had Substantial Error Rates:", with five servers in gray, each with a triangle with an exclamation point inside. The graphic is labeled "3 systems had > 80% document failure to extract rate" and "2 systems had > 95% non-match rate. Prepared by the IDSL for the DHS Science and Technology Directorate.

 

Infographic with three sections, titled "Remote Identity Validation Technology Demonstration - Presentation Attack Detection (PAD)". The first section has three graphics and is titled "FACILITATION: Tested with over 600 consented, bona fide users". The first two graphics are under a heading "Active PAD Subsystems (6 tested): Direct user interaction and leverage smartphone sensors". The first graphic is labeled "Median subsystem validated 85% of bona fide users Best: 94% | Worst: 41%” and shows a person taking a selfie with a smartphone. The smartphone is authenticating the selfie by checking three things, and says “Accept Selfie”. The second graphic is labeled: "More errors for older users Median error rates by age: 18-45: 9%, 46+: 20%". It shows a person holding a smartphone and being confused. The third graphic is titled "Passive PAD Subsystems (15 tested): Classifies images or videos", and shows two servers, one for images and one for videos. The image is labeled "Images were processed faster than videos". Below the graphic, text reads "Median subsystem correctly classified all images/videos of users Best: 100% | Worst 62%". The second section has three graphics and is titled "SECURITY: Tested with over 1,200 presentation attacks. The first one is titled "The median subsystem successfully detected 78% of presentation attacks". Below the graphic is text "Best: 100% | Worst: 0%" and shows a person wearing a mask and hood taking a selfie. The smartphone fails to authenticate the image and says "Detect attack". The second graphic shows a person wearing a mask and hood and holding their hands up in frustration. They are connected to a smartphone and server which is not authenticating them. The heading reads "Both active and passive systems can provide high security." The text below the graphic reads "2 active and 2 passive subsystems stopped all attacks". The third graphic shows a happy person wearing a mask and hood and holding a tablet smart device and a paper photograph, both of which have a picture of the same person. The heading reads "Screen and printout attacks can still succeed on some systems". Below the graphic, text reads "Error rate of up to 88% (median 2%)". The third section is titled "Smartphone Type Affects Performance". Text below reads "Subsystems tested using three smartphones- performance depended on device". A graphic shows three different smartphones. Text below the graphic reads "Facilitation", "Error rate difference up to 14% across devices (median 3%)”, "Security" and "Error rate difference up to 52% across devices (median 7%)”. Prepared by the IDSL for the DHS Science and Technology Directorate.

Results summaries are available at https://mdtf.org/rivtd/Results2023?Length=0, and previously recorded webinars are available at Remote Identity Validation - Maryland Test Facility.

Additional Information

Check the MdTF Remote Identity Validation Technology Demonstration website for the most up-to-date information about upcoming events, or sign-up with peoplescreening@hq.dhs.gov to be included in communication about future events.

Last Updated: 03/05/2025
Was this page helpful?
This page was not helpful because the content